A Neuromorphic Digital Circuit for Neuronal Information Encoding Using Astrocytic Calcium Oscillations

Area of Science:

• Neuroscience
• Computational Neuroscience
• Bio-inspired Computing

Background:

• Astrocytes actively participate in neural information processing by encoding data through intracellular calcium (Ca2+) signal modulation.
• Understanding astrocyte function is crucial for developing next-generation bio-inspired computing systems.

Purpose of the Study:

• To present efficient hardware realizations of astrocytic calcium oscillations and neuronal firing dynamics.
• To enable hardware simulation of astrocyte models for bio-inspired computing applications.

Main Methods:

• Utilized linear approximation and single constant multiplication (SCM) techniques for efficient hardware execution of the nonlinear astrocyte model.
• Implemented a low-cost hardware architecture on FPGA to simulate astrocytic calcium oscillations.
• Employed De Pittà and Integrated & Fire (IF) models for astrocyte and neuron dynamics, respectively.

Main Results:

• The hardware architecture successfully demonstrated essential features of Ca2+ modulation, including amplitude modulation (AM), frequency modulation (FM), and combined amplitude-frequency modulation (AFM).
• Quantitative and qualitative analyses, including phase plane analysis, showed good agreement between MATLAB simulations of original models and FPGA-executed digital circuits.
• The neuromorphic astrocyte circuit successfully demonstrated AM/FM/AFM calcium signaling in real-time operation on FPGA.

Conclusions:

• Developed a novel neuromorphic circuit of astrocytes capable of real-time demonstration of various calcium signaling modulation modes.
• The proposed hardware architecture has potential applications in self-repairing systems and interfacing biological cells with artificial neural networks.